Low-frequency noise study on junction geometry of magnetic tunnel junction (MTJ) sensors for enhancing sensitivity

Low-frequency noise study on junction geometry of magnetic tunnel junction (MTJ) sensors for enhancing sensitivity

Abstract

Detection of ultralow magnetic fields between 1 femtoTesla to about 100 picoTesla has important applications in remote sensing and biomedical imaging such as magnetocardiogram and weapon detection. Currently, ultralow field detection is dominated by relatively large, expensive, power-consuming sensors such as superconducting quantum interference devices (SQUIDS). However, magnetic tunnel junction (MTJS) sensors are very promising for this purpose because they are low-power, low-cost, compact-in-size, compatible with silicon processing, wide operating frequency, and room-temperature operation. MTJS should be able to push magnetoresistive sensor technology into 1 picoTesla sensitivity regime because MTJS have a much larger magnetoresistance (MR) than giant magnetoresistive (GMR) spin valves. However, low-frequency noise is limiting the minimum detectable magnetic field. Previous experimental and theoretical works both indicate that MTJ junction geometry has significant influence on noise. The studies on other spintronic devices including GMR spin valves and La0_7Sro_3MnO3 thin films also show geometry dependence of noise. However, this dependence was not systematically studied in MTJS before. Furthermore, optimization of junction geometry can potentially be an effective means for suppressing MTJ sensor noise.

The purpose of this project is to carry out systematic experimental studies on the influence of junction geometry on MTJ low-frequency (0.1 Hz to 100 kHz) noise. Based on the previous theoretical derivation, we will first predict the relations between noise levels (shot noise, Johnson noise, electronic l/f noise, magnetic l/f noise, and thermal magnetic noise) and MT] junction geometry parameters [l. junction size, 2. junction resistance-area (RA) product, 3. magnetic free-layer thickness]. Then we will experimentally study the influence of junction geometry on MTJ noise and build a MTJ sensor prototype with optimized junction geometry. The preliminary results indicate that this project is feasible and promising. The novelty and uniqueness of this project stem from the unexplored research direction of MT] noise, and the innovative approach for enhancing MTJ sensitivity.

The research outcomes of this project will have great impact in pushing forward the frontier of scientific knowledge on low-frequency noise and advancing spintronic magnetic sensor technology on noise reduction. The mechanisms of electronic l/f noise and magnetic l/f noise have long been an important research topic with many questions remaining unanswered. This study will shed light on the unknown l/f noise mechanism and complete the theoretical model. The research results will also provide clear design guidance for reducing sensor noise level and improving sensor sensitivity by fine-tuning the MT] junction geometry.

(397 words)